Zero-CPU Collection with Direct Telemetry Access

TL;DR

This paper introduces DART, a novel approach enabling switches to directly insert telemetry data into collectors' memory, significantly reducing CPU load and improving scalability in network telemetry collection.

Contribution

DART is a probabilistic method allowing switches to write telemetry data directly into collector memory without coordination, enhancing scalability and reducing CPU overhead.

Findings

Achieves 99.9% query success rate with minimal CPU involvement.

Uses approximately 300 bytes per flow for telemetry data.

Prototypes on commodity hardware demonstrate high effectiveness.

Abstract

Programmable switches are driving a massive increase in fine-grained measurements. This puts significant pressure on telemetry collectors that have to process reports from many switches. Past research acknowledged this problem by either improving collectors' stack performance or by limiting the amount of data sent from switches. In this paper, we take a different and radical approach: switches are responsible for directly inserting queryable telemetry data into the collectors' memory, bypassing their CPU, and thereby improving their collection scalability. We propose to use a method we call \emph{direct telemetry access}, where switches jointly write telemetry reports directly into the same collector's memory region, without coordination. Our solution, DART, is probabilistic, trading memory redundancy and query success probability for CPU resources at collectors. We prototype DART using commodity hardware such as P4 switches and RDMA NICs and show that we get high query success rates with a reasonable memory overhead. For example, we can collect INT path tracing information on a fat tree topology without a collector's CPU involvement while achieving 99.9\% query success probability and using just 300 bytes per flow.